Time Sensitive Control of Aerospace Operations

Abstract : This research is motivated by an effort to integrate temporal and spatial aspects of a large-scale discrete event system (DES) for dynamic control (i.e., planning and re-planning) purposes. The dynamic planning and re-planning problems require satisfaction of two types of constraints, namely temporal and spatial, and the challenge here is to integrate both these constraints in a single analytical formulation of the problem. In addition, the constraints may not be all quantitatively specified; therefore, there is a need to have provisions for both quantitative and qualitative constraint-handling in this formulation. The dynamic control of Discrete- Event Systems (DES) often requires revising a produced temporal model during and/or after system specification phase, e.g., the constraints or system/mission requirements may change during or before a plan's execution. This research resulted in extensions to the existing point interval logic (PIL) that addresses these challenges. The extension allows for a larger class of temporal systems to be handled by incorporating an enhanced input lexicon, representation of flexibility in temporal specifications, an improved verification and inference mechanism, and a suite of analysis tools. The PIL formalism has been shown to incorporate temporal and spatial information separately from each other. The present implementation offers a toolkit of graph-based algorithms for implementing inference and verification mechanisms, revision algorithms for identifying the type of change before making the revision to the plan, and the application of the logic for temporal and spatial knowledge representation and masoning. A recent attempt on combining the graph-based revision and the inference mechanisms into an algorithm that can efficiently handle the dynamic change has yielded promising results.